惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

博客园 - 三生石上(FineUI控件)
月光博客
月光博客
S
SegmentFault 最新的问题
有赞技术团队
有赞技术团队
Stack Overflow Blog
Stack Overflow Blog
Engineering at Meta
Engineering at Meta
T
The Blog of Author Tim Ferriss
The GitHub Blog
The GitHub Blog
小众软件
小众软件
Hugging Face - Blog
Hugging Face - Blog
IT之家
IT之家
宝玉的分享
宝玉的分享
A
About on SuperTechFans
Vercel News
Vercel News
P
Proofpoint News Feed
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
博客园 - 【当耐特】
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
V
Visual Studio Blog
Jina AI
Jina AI
Y
Y Combinator Blog
T
Tailwind CSS Blog
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
Last Week in AI
Last Week in AI

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
Supply Chain Attacks Aren't Just a Big Library Problem — ...
LuzAramburo · 2026-05-19 · via DEV Community
Cover image for Supply Chain Attacks Aren't Just a Big Library Problem — Here's What You Can Do Today

LuzAramburo

In May 2026, a worm called Shai-Hulud compromised 42 TanStack packages — including @tanstack/react-router, a library sitting in millions of JavaScript projects. It was live for about 3 hours. That was enough. If you installed dependencies that day, you may have been affected without knowing it. This post isn't for the people who maintain those libraries. It's for the rest of us — the developers who just use them.

"Fun fact" 1
It was live ~3 hours. @tanstack/react-router alone gets 12.7 million weekly downloads. Meaning that it had ~225K downloads in the ~3 hour window — just for react-router. That's one package. The attack hit 42 @tanstack/* packages total.

Supply chain attacks used to feel like someone else's problem. A big library gets compromised, the maintainers fix it, life goes on. The Shai-Hulud worm changed that framing. It spread automatically to every package its victims maintained, turning regular developers into unwilling distributors of malware. Here's what happened, and what you can do today to reduce your exposure.

What happened?

  1. The workflow used pull_request_target, which runs with the base repo's trusted permissions — including access to secrets and a build cache shared with the real release pipeline.

  2. But it also checked out and executed the fork's code for benchmarking purposes. That's the dangerous mix: a stranger's code running with your own repo's trust.

  3. The attacker didn't need to steal anything immediately. They just poisoned the shared cache and waited. Hours later, the legitimate release pipeline ran, picked up the tampered cache without knowing it, and published the malicious packages itself — using TanStack's own valid credentials.

The key insight: the misconfiguration wasn't obvious. The benchmarking intent was reasonable; the mistake was not realizing that pull_request_target + "run the PR's code" is always a dangerous combination regardless of what you're trying to do with it.

"Fun fact" 2
The worm had a dead-man's switch. It planted a background service that polled api.github.com/user with the stolen GitHub token every 60 seconds. If the token was revoked — meaning GitHub returned a 40x response — the service triggered rm -rf ~/, wiping the user's entire home directory. You had to disable and remove the monitor service before revoking any credentials

The good news

Th pull_request_target Pwn Request specifically requires a public repo where strangers can open PRs.

The bad news

That said, the other parts of this attack (cache poisoning between workflows, OIDC token over-scoping) can still apply to private repos if your GitHub Actions workflows have similar misconfigurations

So far this sounds like a problem for library maintainers. It isn't. You don't need to maintain a library with millions of downloads to be exposed. You just need to run npm install on the wrong day at the wrong time. The moment a compromised package lands in your node_modules, you're part of the chain too.


How to prevent similar incidents?

min-release-age & ignore-scripts

NPM — .npmrc

min-release-age=7
ignore-scripts=true

Enter fullscreen mode Exit fullscreen mode

  • min-release-age=7 blocks packages published less than 7 days ago
  • ignore-scripts=true prevents lifecycle scripts like preinstall/prepare from running on install — which is exactly the vector the malicious optionalDependency used.
  • ⚠️ npm CLI v11 is required. Upgrade to Node 24, or manually install npm v11.

pnpm (10.16) — pnpm-workspace.yaml

minimumReleaseAge: 10080  # minutes — 7 days

Enter fullscreen mode Exit fullscreen mode

  • minimumReleaseAge is on by default, it uses 1 day
  • Requires version 10.16 to avoid a bug that ignored this config.
  • ignore-scripts: pnpm v10 stopped running preinstall/postinstall scripts by default

Yarn Classic

  • min-release-agenot available. No equivalent exists in v1.
  • yarn install --ignore-scripts exists but only as a CLI argument

Yarn Berry (v2+) — .yarnrc.yml

npmMinimalAgeGate: 10080   # minutes — 7 days
enableScripts: false        # ignore-scripts equivalent

Enter fullscreen mode Exit fullscreen mode

  • ⚠️ Requires Yarn 4.10 and applies globally with no way to scope it

Dockerfile

ENV npm_config_min_release_age=3

Enter fullscreen mode Exit fullscreen mode

  • npm ci does respect .npmrc settings and environment variables — the npm_config_* convention works for both install and ci.
  • ⚠️ Still requires node v24 or npm v11, otherwhise ENV npm_config_min_release_age=3 in your Dockerfile will be silently ignored

Option A — Upgrade to Node 24 in Docker only

FROM node:24-alpine
# npm v11 is included, min-release-age works
ENV npm_config_min_release_age=3

Enter fullscreen mode Exit fullscreen mode

Option B — Stay on your Node version, manually install npm v11

FROM node:22-alpine
RUN npm install -g npm@11
ENV npm_config_min_release_age=3

Enter fullscreen mode Exit fullscreen mode


Sources